convnext-tiny-new-5e-5
This model is a fine-tuned version of facebook/convnextv2-tiny-22k-384 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.2310
- Accuracy: 0.9405
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.7921 | 1.0 | 1099 | 0.4486 | 0.8680 |
| 0.6062 | 2.0 | 2198 | 0.3753 | 0.8926 |
| 0.4537 | 3.0 | 3297 | 0.3196 | 0.9113 |
| 0.4313 | 4.0 | 4396 | 0.2750 | 0.9241 |
| 0.3747 | 5.0 | 5495 | 0.2922 | 0.9201 |
| 0.3382 | 6.0 | 6594 | 0.2780 | 0.9276 |
| 0.3065 | 7.0 | 7693 | 0.2569 | 0.9292 |
| 0.2331 | 8.0 | 8792 | 0.2539 | 0.9344 |
| 0.2127 | 9.0 | 9891 | 0.2480 | 0.9352 |
| 0.2409 | 10.0 | 10990 | 0.2452 | 0.9344 |
Framework versions
- Transformers 4.39.3
- Pytorch 2.1.2
- Datasets 2.18.0
- Tokenizers 0.15.2